fixed vs. cathode bias

[TUBESMAN]

I have looked at the data sheet of for the EL34. On one page it says if you triode strap a pair in push pull, using cathode bias, you can obtain 14W with a 10k a-a load. Ok, I dont like a 10W(+) cathode resistor for bias.

What changes do I need to do too achive similar performance with fixed bias?

I'm most interested in maintaining the load impeadence at about the same value. But mostly it be best if the distortion could be kept at the %0.3 level claimed for this mode of use. More power would be nice.

This is a general interest question because I see such wide differences in behavior from fixed to cathode bias. I'm curious how to tweek the settings from one to the other.

[JohnAtwood]

There are multiple dimensions to this issue. The most noticeable one is that in the data sheets, the output power for cathode bias is usually less than for fixed bias. The main reason is that the voltage drop across the cathode resistor takes away from the plate voltage available across the tube which gives less power output. Thus, if you want to run a tube with fixed bias using cathode-bias specs, drop the plate supply voltage so that matches what the tube will see when run at the cathode-bias specs. The fixed-bias voltage will be a negative voltage that is the same as the voltage across the cathode resistor.

Another factor is that push-pull amplifiers using cathode bias don't handle being driven hard very well, since as the average plate current increases, the cathode bias increases, tending to cut-off the tube. On the other hand, with fixed bias, a push-pull stage can typically be driven harder into class AB1 and even AB2 (grid current) with no shift in the bias point. In practice, this means that for a given tube and plate supply voltage, you can get higher power at lower distortion with fixed bias, even ignoring the extra drop across the cathode resistor. The bias shift that happens when a push-pull cathode-biased amp is driven hard has a time constant that determines how fast it recovers, typically a fraction of a second. This can make cathode-biased amps sound "woozy" when handling large signals.

With triode power amps, both push-pull and single-ended, there is an "optimum" load impedance that gives the most power for a given distortion. This is usually what is specified in the data sheets. If the load resistance is lowered, the distortion rapidly increases. If the load resistance is increased, the maximum power available slowly decreases, but the distortion also decreases. Thus is is usually better to err on having the load resistance on the high side of optimum than on the low side. The "optimum" load resistance is quite dependent on the power supply voltage, with an increase in power supply voltage increasing the optimum load resistance.

I hope this helps. You might want to check out the chapters in the RDH4 (Radiotron Designer's Handbook, 4th edition) on output stages for more discussion on this.

- John Atwood

[TUBESMAN]

what you say makes perfect sense. I figured some of it myself but having looked at lots of data sheets, I did not see a patern to confirm my suspicions. I'll look at RDH and see if I can find more information.

[revintage]

In your example the triode-strapped EL34s works in Class-A if fed with B+ of ca 440V and should normally have an unbypassed(also try bypassed and judge by listening) common resistor. This is also the working point for the classic Williamson amp.

Don´t worry about the resistor! If so, buy a hefty one made for heatsink mounting.

In the Class-A case cathode bias is recommended. With a good OPT, single stage driver(why not with transformer PI) and no NFB this will make a nice sounding amp. Next to DHT........

If going for lower Raa and AB the resistor should be bypassed or fixed bias be used as John explained.

Also check the simple Excel-application I have made:

http://www.revintage.se/PPABAMP.xls

It can be used for both triodes and pentodes.

[TUBESMAN]

I'll look into some of these things but I don't feel transformer PI and no NFB are capable of producing an accurate amplifier. Its possible the distortion may be pleasant sounding, but it was not part of the original input signal so it doesnt belong in the output. I also need to have close to zero DC offset in the OPT primary so cathode bias may be a problem.

[TUBESMAN]

I like the spreadsheet, thanks for sharing your knowledge

[revintage]

I'll look into some of these things but I don't feel transformer PI and no NFB are capable of producing an accurate amplifier. Its possible the distortion may be pleasant sounding, but it was not part of the original input signal so it doesnt belong in the output. I also need to have close to zero DC offset in the OPT primary so cathode bias may be a problem.

Hi,
Glad you liked the spreadsheet.

Many people consider NFB being part of what makes amp sound worse.

Remember, a great deal of the hiend market amps are built up the way I recommended ;-). But then using 2A3, 300B etc. No NFB!

If you must have zero offset you have to check the threads where we use mainstoroids as OPTs. You can use a current mirrors and CCS (or two CCS) in the cathode circuit to ensure zero offset!

[TUBESMAN]

I suppose it is as "biased" to fault NFB for everything bad in sound as it is to say that it is without fault. Lets just say that a bad design with or without NFB will have problems.
As far as the "High End" (finger down throat) here is my assessment of it.

http://www.biline.ca/critic4.htm[/url]

Trust me, if you can hear it, it can be measured. And if you cannot measure it yet, then devise a test for it. If it is a new phenomenon, then you will get a special award in Oslo Norway. That is what science is.

Personal taste in sound is ok. But be honest, it is personal taste that one form of distortion or another "sounds better" than another . The purpose of a power amp is to 1: raise the output voltage level. 2: lower its impedance so it can deliver current without influencing the frequency response. Anything else it does is either distortion or personal taste.
This link looks at the pecking order we sometimes choose to ignore in our persuit of I don't know. I was going to say Fidelity, but it is obvious that is not a consideration any longer.

http://www.biline.ca/critic1.htm

My CO2 extinguisher is ready.
All I am saying is know the difference between taste in sound and the original reason for High Fidelity: to recreate the original performance in your room. If you need 10% distortion added to give you that experience, OK, add it on your own with a device that can be turned off or turned down. I remember the old days when Color TV was new. Some companies sold multi color striped acetate film that you placed over your BW TV screen so you too had "Color TV". People bought them. Just like SE tube audio.
I do tubes because I grew up with them. I can do an amp with parts that I can still see without a magnifier. But, I will not compromise performance.

[revintage]

Tubesman,

Maybe you thought I was picking on you but that was wrong!

SorryI used "hiend market amps", I ment good, music reproducing amps.

I just reflected on your first post where you mention the low THD of a PP Class-A triode EL34. With the help of a driver that also gives distortion cancellation(not local NFB or NFB at all) you can actually get below .1% THD up to 10W w/o NFB.

To be honest, do you really think your reference is to be looked upon as a serious reference? So much text without mentioning MUSIC at all .

Will not discuss this matter anymore, but will be happy to discuss technical parts of your interesting upcoming amplifier-project . Good luck!

[Sheldon]

Originally posted by TUBESMAN
the original reason for High Fidelity: to recreate the original performance in your room.

That's impossible with typical hardware in any ordinary home listening room. What we can try to do is to create the most effective illusion (or the goal could be the most pleasing illusion). The human perceptual apparatus cannot be ignored as a major part of the illusion. Absolute accuracy in the electronic chain may not hurt, but is not necessary either.

Sheldon